THE PESPAfiATION OF THIN FILMS OF InSb
ON GKTSTALLINE SUBSTRATES
B. S., Park College, 1959
A MAbTER»S THESIS
suhmitted in partial fulfillment of the
requirements for the degree
MASTER OF SCIENCE
department of Pl^aics
KANSAS STATE UNIVEESTTI
^^"0- TABLE OF COKTaiTS
E3JIPM;irT .\ND PftOCJEDUHE g
KESULTS AND CONCLUSIONS 2i
Studies of the properties of thin films of indium entimonide have been
performed using polycrystaUlne films. (3), (6). The interpretation of
experimental data from such polycrystaUlne films is complicated by effects
which are due to structural monuniformity. It would therefore be of value
to be able to study monocrystalline films. The ultimate goal of the work
described in this paper was the production of monocrystalline films of InSb
suitable for further studies of the properties of thin films,
A promising and widely \ised method for preparing monocrystalline filmg
is epitaxial growth. This method is based on the fact that films deposited
from the vapor phase onto suitable crystalline substrates often are composed
of large oriented crystallites.
The phenomenon of one crystal grown upon another with some definite
and unique orientation relationship between their 'crystal axes was first
observed in the form of natural growths among minerals. The term epitaxy
denotes this phenomenon. Many naturally occurring cases of oriented inter-
growth have been observed and comprehensive lists of the known examples have
been compiled, (7).
Observation of these intergrowths led to attempts to obtain oriented
intergrowths in laboratory experiments. The parallel growth of sodium
nitrate frcm solution <m the surface of a calcite crystal by Frankenhelm
in 1836, appears to have been the first successful attempt. The first
systematic experiments were performed by Barker who sttdied the growth from
solution of crystals with related structures, such as the alkali halideil^
upon each other. He found that some alkali halides oriented upon each othw
while others did not, and concluded that oriented growth is more likely to '
occur if the molecular volumes of the two alkaU halides are nearly equal. (2).
After the discovery and development of x-ray diffraction and the conse-
quent tramendous increase in knowledge of crystal structure, fioyer repeated
and extended the work done by Barker, On the basis of his results he devel-
oped a theoiy of epitaxy. This theory centered on the necessity of a paral-
lelism of the lattices accompanied by a aaall misfit value for the lattices,
Electron diffraction was discovered about the time Jbyer was doing his
work. The application of electron diffraction to the study of epitaxy
greatly increased the possibilities for study, making it possible to study
much thinner growths and deposits other than those grown from solution.
With the application of electron diffraction to work on various types of
deposits, it was found that epitaxy occurs in such varied deposits as chem-
ically gxx>\m layers, electrodepo sited metal films, and metal layers condensed
from the vapor phase. Van der Merwe compiled a comprehensive summary of
known cases of epitaxy and Paahley has compiled a detailed survey of those
cases which have been studied using electron diffraction techniques. (7), (12),
The investigation of epitaxy of metals upon non-metals has been carried
out almost exclusively by electron diffraction. Lassen's discovexy of silver
in parallel orientation when deposited from the vapor phase on a rocksalt
cleavage face was the first known occurrence of this type. (7),
Lassen and Bruck reported the growth of good, single ciystal, thin
films of silver by deposition on heated rocksalt substrates. They reported
the films grew in parallel orientation, which corresponds to a misfit value
of -ZTi. itoyer challenged this observation on the basis that other possible
orientations would result in much aaaller misfits and would therefore be the
ones expected to occur. Lassen and Bruck, while agreeing one might expect
other orientations on the basis of misfit, confirmed that it was neverthe-
less the parallel orientation which occurred, (9),
Bruck and Rudiger, in their independent studies of metals deposited on
rocksalt, fluor, and caloite, found that orientation of the films was markedly
dependent on the substrate temperature during deposition, i'hen the deposi-
tion was made onto substrates held at temperatures above a certain critical
temperature (the epitaxial temperature) well oriented films were obtained.
For depositions made on substrates at temperntures below the epitaxial tem-
perature there was not complete orientation, ^he epitaxial temperature was
found to be different for different metals, (2), (7),
aiirari. carried out many experiments with metals on non-metals and
found that in some cases where the substrates were preheated to several
hundred degrees Centigrade and then cooled to some lower temperature for
the deposition there was an improvement in the degree of orientation achieved,
Oriented growths of metals upon mica, calcite, fluorspar, and mineral
sulphides have been studied extensively. These are characterized, in general,
by the occurrence of mixtures of orientations which are temperature dependent.
The misfits found in these films are often voiy high, (7).
More recently, Kehoe examined the growth process of silver films by
continuously monitoring the diffraction patterns from films being deposited
in the diffraction camera. He found that good orientation occurred at temper-
atures appreciably below previously reported epitaxial temperatures, Kehoe,
however, used much slower rates of deposition than had been used previously.
Sloope and Tiller, working primarily with silver on rocksalt, were the
first to perform ^stematic experliaents to establish some of the iaportant
factors for the production of good single crystal thin films by epitaxial
growth. They found the most perfect films, structurally, were produced on
substrates that were given a preliminary heat treatment to reduce the amount
of absorbed gao and other foreigj^ matter on the surface, 4iort periods of
preheating were found to bo as effective as extended periods. It was also
noted that if the substrates were preheated to temperatures where thermaJ.
etching took place the resultant films were porous. In studying the relation
between the substrate temperature and deposition rate they found that there
is actually no epitaxial temperature, but that the required substrate temper^
ature varies with the rate of deposition, (10), *
Kalinkin, Alesicovskii, oergeeva, and ^itraknow hava obtained monociystal-
line films of CdSe on surfaces of NaCl, KOI, and KBr, On the basis of their
studies they concluded that the orientation of the films did not depend to
any significant degree on the surface relief of the substrate and that in
obtaining monociyatalline films the most important parameters were the
substrate temperature during deposition and the rate of deposition. They
also observed that a preliminary heating of the cxystal substre\te increased
the ability of the substrate to cause orientation of the film, (A).
Extensive studies have been made of Uie occxirrence of and conditions
for epitaxy, but there still does not exist a good theory of epitaxy to provide
guideposts for work with previously untried materials.
There have been many theoretical considerations of epitaxy but the
majority of them were based on the concept of a good georaetrical fit between
the lattices of the substrate and the overgrowth and completely ipiled to
explain the many large misfits and varied orientations that have bean observed.
It now seems to be commonly agreed that a greater knowlc-dge of the process of
nucleation and of surface forces must be obtained before any comprehensive
theory of epitaxy can be evolved.
E.5UIPMENT AND PK)CEDUJaS
Plate I shows the vacuum chamber arrangement used throughout this inves-
tigation. All evaporations ¥*ere joade inside a 3 inch glass tee. The tee
was seided to a brass flange on the throat of the pumping systen by iaeans of
a soft rubber gasket. The pumping system used to evacuate the chajober con-
sisted of a chain of two oil diffusion pumps and a mechanical forepurap and
was capable of an ultimate vacuum of 10"^mm Hg, The main pumping system
could be sealed off just ahead of the chamber and the chamber itself roughed
down by a second mechanical pump after it had been opened to the atmosphere.
Thus the pumps were never turned off or exposed diwctly to the air. The
gas pressure in the chamber was measured with a Consolidated Vacuum Corpor-
ation, Philips Gauge type FHG - 09.
Bulk Ina> was evaporated from a molybdenum boat which was mounted on
copper high current leads. These leads passed out of the chamber through
a brass end-piece on the side am of the tee. The power for the boat was
supplied by a 3 kva, U to 1, stepdown transformer regulated by means of a
100 amp Transtat Voltage Begulator in the primaiy.
The brass end-pieces on the top and side arm were sealed to the glasa
by means of Teflon gaskets, the pressure being applied by bolts passing
between opposliig aluminum flanges. The top piece supported the substrate
holder and had the heater and thermocouple leads passing through it. The
substrate holder held both the substrate material and a glass optical flat
approximately 11cm above the bottom of the boat. The substrate was sand-
wiched between the substrate hater in contact vdth the rear surface and a
brass mask over the face. The optical flat also was covered by a brasa
mask having a rectangular slit so that the film deposited on it had two.
long, parallel edges. The tiiLckness meaauromenfcs were made on this film by
an optical method employing interference by reflection of sodium light with
a wavelength of 589OA, The fringe shifts were measured with a telescope
equipped with a microneter head.
The optical flat was not backed with a heater and therefore remained at
a different temperature than was being used for the substrate during the
depod.tion of the film. The actual thickness of the films deposited on the
ciystalline substrntes may have been somewhat different from those measured
on the optical flat due to re-evaporation from the heated substrate. The
relative rate of deposition for the various films was of more interest than
the actual thickness of the film in this study. Thi.s was calculated from
the thickness measurement and length of time to evaporate the charge. The
deposition rate, as measured, represents only an average rate since the actxial
deposition rate varied somewhat throughout the deposition due to factors such
as the reduction of the size of the charge in evaporating and temperature
changes of the crucible.
The substrate heater was a ceramic slab on one side of which ohromel wire
was threaded through thin parallel cuts and cemented into plaoe with Eccooeram-
CS ceramic bonding and sealing confound. The other side was lapped flat to
provide thermal contact with the substrates. Power was supplied to the
heater fi%^ a transfomer regulated by a variac.
The substrates were freshly cleaved from bulk samples of C;aF2, KBr, LiF,
and NaCl, The only exception to this was a thermally etched (111) surface of
Nad which had been exposed to the air for an extended length of time, Caf^
was used because it cleaves easily along the (HI) plane. The (111) plans
was thought to be the most likely to produce epitaxy since InSb has the
zinc blende structure and if it grew on a (111) face in parallel orientation
KiPLAHATION OF PUTE I
CttUitajr view of vaetuia ehaober axmngemcnt*
Subatrate heater £• Optical flat
^8trat« F, Braaa oaak
Bl«M wtLwk Q» Ii^ybdenum boat
SiibstrBte holder H« Current leads
the crystal lattic* would grow as alternate parallel layers of In and Sb,
KBr was tried because it has a lattice constant nearly the same as that of
InSb, NlaCl was tried because most of the good single crystal thin films of
yarious metala have been grown on rocksnlt substrates. LiF was used because
it was available in large pieces with good crystal structvire.
After the substrate was cleaved, a copper-constantan thermocouple waa
attached to its surface with Eccobond 58C solder and the substrate was heated
approximately 1 hour to dry the solder. The substrate was then immediately
mounted in the vacuim system. Prior to evaporation the substrates were
heated in vacuo to a temperature somewhat higher (50°C to 200°C) than the
intended substrate temperature during deposition, and were held at that tem-
perature about 2 hours or more. The substrate was then cooled in a few
minutes to the desired substrate temperature and evaporation was begun. The
pressure in the chamber never rose about IC^mm Hg during deposition of the
films. The substrates were allowed to cool slowly to room temperature imme-
diately upon completion of the deposition. The substrate temperature was
determined from the thermocouple otf measvired with a Rubicon potentiometer
model No. 2700,
anall pieces of substrate with the overgrowth of InSb were carefully
cleaved from the rest of the substrate for study by reflection electron
diffraction. In one situation it became necessary to examine the film by
transmission diffraction. In several cases pieces of the substrate not
coverwcv oj/ uhe In2b were also cleaved off for examination by electron diffrac-
tion. The electron diffraction patterns were obtained in a RCA model EMU-2D
electron microscope equipped with a reflection diffraction attachment.
After examination by electron diffraction, Pt-Pd shadowed carbon surface
replicas were made of the film surfacRS. The shadowing was done by evapor-
ating th6 t>t-Pd wire from approjdmately 15° with respect to th ^ film surface.
The carbon was then deposited at normal incidence. The replicas were rexaoved
in dilute nitric acid and mounted on grids for exajnination in the electron
BESULTS AND CONCLUSIONS
Fifteen usable InSb films were deposited on (111) cleavage faces of
CaFg. Hiese were deposited -with the substrates at various tempera ttirea
ratgiag txom. 25°C to 300°C amd at various rates of depositioa raaglag from
200A/min to lOOOA/min. -Vhen examined by electron diffraction all of these
films gave good InSb diffraction patterns, but there were significant vari-
ations in the types of pattei-ns and in the appearance of pure In and pur«
Sb lines in the patterns. The pertinent data for the films are compiled in
Slower deposition rates were found to be more favorable for the foma-
tion of InSb films with larger crystallites. Films A-11 and A-3 were both
deposited on substrates held at 200°^, but the deposition rate for film A-.3
was double that of A-11. The diffraction pattern for film A-11 (Plate II,
Fig. 1) shows spots and grainy rings ■whereas the diffraction pattern for
film A-3 (Plate II, Fig. 2) shows only continuous rings. Films A-I4 and A-I5
were also deposited on substr-tes held at the same temperature (300°G), but
the depo^tion rate for film A-I5 was double thr.t of film A-I4. Here, as
before, the diffraction pattern from the mre slowljr deposited film, film
A-14 (Plate III, Fig. 1), contained well-defined spots while the pattern
from the more rapidly deposited film, film A-I5 (Plate III, Fig. 2), con-
tained only continuous rings. The spotted patterns from the films deposited
with the slower deposition rates show the crystallites in these films are
somewhat larger than those iii the corresponding films formed at faster
deposition ra.es and having only continuous rings in their patterns.
There was a definite darkening of segments of the rings in the diffrac-
tion pattern from film A-4 (Plate IV, Fig. 1) deposited at a substrate
Data for InSb
^ Diffraction Pattern
(d-spacings for lines present)
InSbi 3.74,2.29,l«95,1.49,1.32j In: 1.68l ^: 3.10,1,77
3.74,2.29,1.95,1.62,1.49,1.145; In: -^71} 3.10
2.29,1.95,1.62,1,49,1.32; Sb: 3.10} Unknown: 3.46
0,907 '-eak spots
InSb: 3.74,2. 29,1.95,l»62,l»^9,1.32,l,25,1.145,0,907j
Sb: 3,10 Weak spots
2,29,1,95 ,1.49,1.32,1.U5 J In: 2.72,1,68,1.47
3,74,2,29,1.95,1.49,1.32j In: 2.72,1,631 ^: 3.10,1,76
2.29,1.95,1.145 Veiy faint pattern
2,29,1.95,1.62,1.49,1.32,1.25,1.145; In: 2.72
InSb: 2.29,1.95,1.62,1.49,1.32,l,25,1.145,1.02j In:272
Sb: 3.54 %)ots
2.29,1.95,1.49,1.32,l.U5j In: 2,72,1.40j Sb: 1,76,
3,74,2.29,1.95 Trananis&ion pattern
3.74,2.29,1.95 ,1.62,1.49,1.32,1.25 ,1.U5,1.09,1. 02,
2,29,1,95 ,1.U5 J In: 2,72,1,685 1,55
* A denotes CaF- substrate; G, LiF substrate; D, KBr substrafce;£, NaCl substrate
EXPLANATION OF PLATE II
Fig. 1, Electron diffraction pattern from film A-11,
Fig, 2. Electron diffraction pattern from film A-3.
EXFUNATION OF PLATE in
Pig. 1. Electron diffraction pattern from film A-U.
Fig, 2, iilectron diffraction pattern from film A-I5,
EXPLANATION OF PUTE IV
Fig. 1, Electron diffraction pattern from Him A-4.
Fig, 2. Electron diffraction pattern from film A-14.
temperature of 25 G, This was interpreted as due to a sli^t tendency towards
orientation of the ciystallites composing the film. For materials producing
monociystalline films, a higher degree of orientation of the crystallites
la found in films deposited at higher substrate temperatures. The only
evidence for orientation of the ciystallites in InSb films deposited at sub-
strate ten^eratures above 25°C was from film A-U, deposited at 300°C. The
aajor spot pattern of multiple spots in the pattern from one setting of film
A-U (Plate III, Fig. 1) suggests that several crystallites were in nearly
the same orientation. This orientation was not characteristic of this entire
film, however, as diffraction patterns from other areas (Plate IV, Fig. 2)
did not show evidence of ciystallite orientation.
In general, the best olycrystalline InSb films produced on CaF2 sub-
strates were produced on substrates held between 200°C and 300**C during
deposition. This is evidenced by the good In* diffraction patterns obtained
from these films with notably fewer pure In and pure Sb lines occurilng in
the patterns. A very excellent InSb diffraction pattern was obtained from
film A-13 (Plate V, Fig, 1) deposited with a substrate temperature of 250°C,
The peculiar background associated with the pattern from film A-13 v>as
also found in the diffraction patterns of film A-? (Plate V, Fig. 2). The
background changed in orientation with different settings of the film with
respect to the electron beam but remained of the same general nature for
all orientations of the samples. No reference to this phenomenon has been
found in the literature and nothing was observed which would possibly explain
its existence in the patterns from these films.
All films deposited on CaF^ substrates at temperatures above 200°C were
extremely loose on the substrates when removed from the vacuum system. In
80IM instances the whole film or a portion of it fell off the substrate
SXPLASATION OF fU!^^ V
1, ELectroft diffraction pattern, from film A-13.
fig, Zt Electron diffraction nattem from film A-8.
during subsequent handling. This looseness was attributed to the difference
in the coefficients of expansion for the two materials.
lii several cases good electron diffraction patterns nere obtained from
the substrate material of the samoles before serious charging effects set in.
These patterns, obtained from the substrate material of films A-7 (Plate VI,
Fig. 1 and Fig. 2), k-lU (Plate 711, Fig. 1), A-l^ (Plate VII, Fig. 2), and
A-17 (Plate VIII, Fig, l), show the substrate material had the very good
crystal structure desired for epitaxy. This is evidenced by the good spot
patterns and Kichuchi lines obtained from these samples. The ^mmetrioal
elongation of the spots in the patterns is characteristic of a lack of pene-
tration of the electron beam into the surface with a corresponding relax-
ation of the Bragg condition.
For a given set of conditions, GaFg seemed to foster better films of
InSb than the other substrate materials used. Diffraction patterns were
obtained from InSb films deposited on (100) faces of KBr and NaCl at 150**C,
on a (100) face of KBr and a (111) face of NaCl at 200°C, and from one film
on a (100) face of LiF p.t unknown substrate temperature. Some of these pat-
terns were very faint. No evidence of orientation or crystallite growth was
found in these films. A much greater separation of components, giving rise
to pure In and pure 3b lines in the diffraction patterns, was found to exist
in the films on KBr and LiF than in comparable films on CaF2. Good InSb pat-
terns were obtained from the films on NaCl.
The surface micrographs (see Plate VIII, Fig, 2, and Plate IX, Fig. 1
and Fig, 2, for representative examples) were prepared with the idea that
some information as to crystallite size, shape, or orientation ai^t be
obtained but such was not the case, A correlation between the appearance of
the film surface and its diffraction pattern was sought, but there was no
EXPLANATION OF PLATE VI
Fig. 1, Electron diffraction pattern from the substrate
of film A-?.
Fig, 2. iilectron diffraction pattern from the substrate
of film A-7.
EXPUNATION OF PLATE VII
Fig, 1, Electron diffraction pattern from the substrate
of film A-14.
Fig, 2, Electron diffraction pattern from the substrate
of fiLa A-16.
EXPLANATION OF PLATE VIII
Pig, 1, Electron diffraction pattern from ttie substrate
of film A-17.
Fig. 2. fiectron miscroscope laicrograph of a airfsce
replica of film A-9.
EXPUNATION OF PLATE IX
Fig, 1, Electron microscope micrograph of a surface
replica of film A-4,
Flg» 2» -.lectiKin fidci'oscope cdcrograph of a surface
replica of film A-2. -
recognizable difference between the surface features of films giving dif-
fraction spots on their patterns ami those of films giving only rings. The
fflicrographs did show that the films produced were, with one possible excep-
tion, continuous. Much of the value of the airface micrographs was lost
because no way was found to remove the carbon replicas from the CaF2 with-
out serious damage to the replica and thus it could not be determined which
structures observed on the film surfaces were due to substrate surface
features and which were inherent with the films.
It Has shown that slower deposition rates favor the growth of InSb films
with larger ciystallites. The slowest rate of deposition used in this inves-
tigation was 200A/xnin. Slower rates were not used because bulk InSb under^
goes component separation upon evaporation and the more volatile Sb initially
evaporates at a much faster rate than the In, Thus, the first layers cm the
substrate are almost pure 3b and the last layers almost pure In with vaiying
combinations of the two components in between. The slower the evaporation
rate the more pronounced this separation tends to become. This separation
requires that some sort of diffusion mechanism must be operative after the
coa^onents reach the substrate in order to produce good InSb films. The
rebuilding of the Ina) lattice by diffusion of the componftnts is probably not
veiy conducive to formation of extensive ciystallites in the films. Thus
it would be of great interest to attempt epitaxial growth of InSb by use of
the three temperature technique whereby the components are evaporated from
separate crucibles onto a substrate held at some given tonperature, V,ith
this method, deposition rates in the order of a few Angstroms per minute could
be worked with and good InSb layers would be developed at all stages of
growth rather than by some process after deposition is completed. Thus,
films with rather large oriented cxystallites mi^t be obtained.
1. Beeching, R.
Electron diffraction. New York: Chemical Publishing, 1936.
2. Buckley, K. ii,
Cxystal growth. Mew York: John '/.'iley, I95I,
3. Dale, £« B. and Senecal, G.
Annealing effects in evaporated laSb films, J. Appl, Ftyr. 33:2526-2532,
U, Kalinkin, I. P. et al.
Preparation of mono crystalline layers of cadmium selenide, Soviet Plyrsics-
SoUd aate, 3:1922-1927. 1962.
5. Kehoe, R, B.
On the texture of evaporated films, Phil, Mag. 2:455-566. 1957.
6. Kurov, G. A, and Pinsker, Z, G,
Investigation of thin films produced by evaporating indium antinionide
in vacuum, Soviet Physic s-Technical Physics, 3:1958-1963, 1958,
7. Pashley, D, iV,
The study of epitaxy in thin surface films. Adv. Phy. 5:173-240. 1956.
8. Pinaker, Z. G,
Electron diffraction, London: Butterworths Scientific Publicationa,
9. Seifert, H.
Structure and properities of solid surfaces, Chicago: University
10, Sloope, B, W, and Tiller, C. 0,
Fomation conditions and structure of thin epitaxial silver films on
rocksalt, J, Appl. ihy. 32:1331-1336. 1961.
11, Thomson, G, P. and Cochrane, W,
Theory and practice of electron diffraction, Londwi: Macmillan, 1939,
12, van der Merwe, J, H,
liisfitting monolayers and oriented overgrowths, Diec, Fard, Soc,
I wish to thank Dr. E, B, Dale whose encouragement and assistance
contributed so greatly to ity vjork and Dr. E« D. Dragsdorf for his valuable
THE PREPAR4TI0N OF THIN FILUo OF InSb
ON CiSSTALLINE OJBbTfl/.TiiS
B. S,, Park College, 1959
AN ABailiACT OF
A UkrJIER^S THESIS
submitted in partial fulfillment of the
requirements for the degree
MASTER OF SCIENCE
Depai^tment of Physics
KANSAS iJTATE UNIVERSITY
In studies of the properties of thin films of InSb it has been found
that the interpretation of the experimental data has been complicated by
effects due to the polycrystalline nature of the films studied. The pur-
pose of this noxk was to develop methods of preparing monocrjrstalline films
of InSb suitable for further studies of the properties of thin films. The
epitaxial growth of films by deposition from the vapor phase on crystalline
substrates has proved highly successful for the production of monociystalline
films of fflciny materials and thus was the method employed for this study.
The crystalline substrates used were primarily (111) cleavage faces of
Ca?2 but several films were also produced on (100) cleavage faces of NaCl,
KBr and LiF and on one (HI) face of NaCl, These substrates were given a
preliminary heat treatment In vacuum and were allovjed to cool to the desired
Bubetrate tezapei-ature for the deposition of the film. The substrate temper-
atures ranged from 25°C to 300°C and the depositiwa rate ranged from 200A/min
to lOOOA/oin for deposition of the various films, Small pieces of the sub-
strate with the InSb film were cleaved from the rest of the substrate for
examination of the films by reflection electron diffractica. After this
examination a Pt-Pd shadowed carbon surface replica was made of the film
for examination in the electron microscope,
Evidence was found that, othoz- conditions being similar, slower rates
of deposition favor growth of larger crystallites. A sli^t orientation of
the crystaJlites was found in the film produced on CaF2 at 25®C, The onljr
evidence for orientation of ciystallites In films produced on substrates at
high«r temperatures came from one arc? of a film deposited on a substrate
held at 300°C. Study of the diffraction patterns from all of the filee
showed that, in general, the best polycrystalline Jn3b films were obtained
from deposition on CaF^ substrates in Uie temperature range of 200°C to 300OC.
The fldcrographs of the carbon surface replicas showed that the films produced
were, with one possible exception, continuous. There was a great yariation
in the surface characteristics of the films but no correlation was found
between the surface features and the type of diffraction patterns obtained
from the films.